Hire Akhil - NLP Engineer

* Advanced Voice Processing and Analytics with AI. * Developed and deployed ML/DL models to analyze call center voice data, extracting key conversational elements and emotions. Built APIs and generated reports to optimize various use cases including debt collection, tele sales and marketing, audit and compliance. * Processed and enhanced call center voice data through cleaning and up-sampling techniques. * Transcribed, diarized and translated voice data utilizing Azure Cognitive Speech to Text Services and Whisper AI libraries and packages. * Manually annotated data for emotions and key conversational elements. * Trained deep learning and machine learning models to detect intent, emotion, and call checkpoints using MFCCs and other voice and text features. * Utilized NLP techniques to analyze text. * Developed use cases for debt collection, verification, and financial transactions. * Fine-tuned a pre-trained call summarization model to condense lengthy calls into concise paragraphs for our specific use case. * Generated reports and developed APIs using Django Rest Framework. * Deployed models on Oracle Cloud Infrastructure and collaborated with our front-end developer for integration.

* 1. P2P Lending Credit Risk Model Initiative * As a co-team lead, I spearheaded the development of a powerful P2P lending credit risk model for our esteemed client, Bondora Capitals. Leading a team of 12, we optimized lending decisions and mitigated default risk. * Coordinated the analysis of borrower data and loan transactions, ensuring precision for data-driven strategies. * Guided the team in building a robust credit risk model using advanced statistical techniques and ML algorithms * Developed a unique multiple-output regression model to predict three target variables accurately. * Identified pivotal risk factors, empowering our client to make well-informed lending decisions. * Designed an end-to-end machine learning pipeline to streamline data processing, model training, and predictive analysis. * 2. ICU Patient Mortality Reduction Initiative * Collaborating with Theorax Dynamics as our valued client, I played a crucial role in reducing ICU-admitted patients' mortality rates by identifying predictive factors. * Conducted thorough EDA on patient data, identifying key patterns. * Collaborated with the data team to clean and transform data, ensuring high-quality analysis. * Developed a highly accurate model (95% accuracy) using Python, Statistics, and Scikit-Learn. * Identified critical factors for patient fatality, leading to preemptive diagnostics for early detection and preventive actions.